Opti welcomes you to the wonderful world of slop readers.
Today as an example: I thought I found a logic error in my (non-botted) code this morning while writing a little guide, so I've been reading slop from 3 different bots all trying to figure it out for about a normal person's workday worth of time by now, to a point that I actually needed a nap in the middle because it gave me an oppressive headache. Fixed it now. And it was an actual logic error (throwing the wrong exception type in a single location)
Faster if I'd just fix the one line, but better, no, or at least, not while there is a risk that I am wrong about the scope. After all it's my own initial mistake so I need a "second opinion", which means 2-3 bots (of different models/weigths) assessing in parallel, or requesting a meat opinion, which (a) has timezone issues during my mornings and (b) is extremely expensive because you need to read the person in on the exact code.
Do you still phrase critiques as if the author authored it?
deleted by author
deleted by author
What all is involved in refereeing a paper?
deleted by author
Ah, so is this the fabled "peer review"? And do you have a say in whether the thing ultimately gets published? Or is that all on the editor?
deleted by author
Opti welcomes you to the wonderful world of slop readers.
Today as an example: I thought I found a logic error in my (non-botted) code this morning while writing a little guide, so I've been reading slop from 3 different bots all trying to figure it out for about a normal person's workday worth of time by now, to a point that I actually needed a nap in the middle because it gave me an oppressive headache. Fixed it now. And it was an actual logic error (throwing the wrong exception type in a single location)
All day for a -1/+1 diff - I'm such a retard.
deleted by author
Faster if I'd just fix the one line, but better, no, or at least, not while there is a risk that I am wrong about the scope. After all it's my own initial mistake so I need a "second opinion", which means 2-3 bots (of different models/weigths) assessing in parallel, or requesting a meat opinion, which (a) has timezone issues during my mornings and (b) is extremely expensive because you need to read the person in on the exact code.
How did you figure it out?
Was it good/valuable or just slop and bs?
deleted by author
did you see this?
https://twiiit.com/MarkKoyama/status/2102204839581155358
deleted by author